Commit graph

46 commits

Author SHA1 Message Date
Mathieu Duponchelle
ad49ae42f7 docs: mark more types as plugin API 2020-06-23 12:10:19 -04:00
Seungha Yang
9c21923f04 nvcodec: Add H264 stateless codec implementation
Introduce GstH264Decoder based Nvidia H.264 decoder element.
Similar the element factory name of to v4l2 stateless codec,
this element can be configured with factory name "gstnvh264sldec".
Note that "sl" in the name stands for "stateless"

For now, existing nvh264dec covers more profile and formats
(e.g., interlaced stream) than this implementation.
However, this implementation allows us to control lower level
parameters such as decoded picture buffer management and therefore
we can get a chance to improve performance in terms of latency.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1198>
2020-04-24 09:23:10 +00:00
Seungha Yang
b4efdeba11 nvdec: Don't hardcode DPB size
Too many decode surface would waste GPU memory. Also it seems to be
introducing additional latency depending on stream. Since nvcodec
sdk version 9.0, CUVID parser API has been providing the minimum
required number of surface. By using it, we can save GPU memory
and reduce possible latency.
2020-04-09 16:30:58 +09:00
Seungha Yang
22eab50907 nvdec: Add fallback for CUDA/OpenGL interop failure
It happens when local OpenGL context belongs to non-nvidia GPU.
2020-03-19 13:58:09 +09:00
Nirbheek Chauhan
266dc41596 nvcodec: Mark class data as may-be-leaked to quiet the leaks tracer
The class data with the caps in it will be leaked if the element is
registered but never instantiated. There is no way around this. Mark
the caps as such so that the leaks tracer does not warn about it.

This is the same as pad template caps getting leaked, which are also
marked as may-be-leaked. These objects are initialized exactly once,
and are 'global' data.
2020-02-12 00:00:51 +05:30
Nirbheek Chauhan
3ca87d9988 nvcodec: Fix crash in decoder on 32-bit Windows
Same fix as 1a7ea45ffd, but I didn't
test the decoder so I missed that the function pointers here weren't
using the correct calling convention too.
2020-02-06 13:39:52 +00:00
Sebastian Dröge
5b8ff98f96 nvdec: Don't leak template caps when registering elements with old NVIDIA driver 2020-02-05 09:49:20 +00:00
Nicolas Dufresne
d393232bc2 nvdec: Do not map GStreamer discont to CUVid discont
Setting the CUVID_PKT_DISCONTINUITY implies clearing any past information
about the stream in the decoder. The GStreamer discont flag is used for
discontinuity caused by a seek, for first buffer and if a buffer was
dropped. In the first two cases, the parsers and demuxers should ensure we
start from a synchronization point, so it's unlikely that delta will be
matched against the wrong state.

For packet lost, the discontinuity flag will prevent the decoder from doing
any concealment, with a result that ca be much worst visually, or freeze the
playback until an IDR is met. It's better to let the decoder handle that for
us.

Removing this flag, also workaround a but in NVidia parser that makes it
ignore our ENDOFFRAME flag and increase the latency by one frame.
2020-01-25 13:39:03 +00:00
Nicolas Dufresne
a28ce16b3f nvdec: Tell the parser we have complete pictures
This sets the CUVID_PKT_ENDOFPICTURE flag in order to inform the decoder that
we have a complete picture. This should remove one frame latency otherwise
introduce by NVidia parser.
2020-01-25 13:39:03 +00:00
Seungha Yang
49bccf0433 nvcodec: Refactor plugin initialization
Create CUDA context per device, instead of per codec and encoder/decoder.
Allocating CUDA context is heavy operation so we should reuse it
as much as possible.

Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/1130
2019-12-24 08:10:14 +00:00
Seungha Yang
14b9a1cffd nvdec: Add support for mpeg4 video decoding with codec_data
Decoder should handle codec_data of mpeg4 video which includes essential
config data.
2019-09-11 14:00:48 +00:00
Seungha Yang
e3508a4f26 nvdec: Update plugin description and fix typo
Use consistent description with nvenc, and fix typo s/devide/device/g
2019-09-11 15:16:45 +09:00
Marc Leeman
3ef503346a nvcodec: minor spell corrects in log messages 2019-09-10 23:13:17 +00:00
Seungha Yang
fa83f086be nvdec: Check flow return of the only current handle_frame() to fix seeking issue
Due to uncleared last flow, decoding after seek was never possible
(last_ret == GST_FLOW_FLUSHING).
nvdec dose not need to keep track of the previous flow return,
and actually the interest is data/even flow of the current handle_frame().
2019-08-30 11:27:27 +09:00
Seungha Yang
be3a3da829 nvdec: Fallback to system memory if OpenGL context could not support PBO memory
If the environment could not support OpenGL PBO memory, nvdec will do negotiation
with system memory as fallback.
2019-08-30 01:36:46 +09:00
Seungha Yang
069fe93452 nvdec: Add support dynamic output format change
Implementing ::negotiate() method to support runtime output format
change. If downstream was reconfigured, baseclass will invoke
::negotiate() method, and nvdec should update output memory
type depending on downstream caps.
2019-08-30 01:36:46 +09:00
Seungha Yang
39f800c449 nvdec: Re-negotiate whenever output format is changed
Input stream might be silently changed without ::set_format() call.
Since nvdec has internal parser, nvdec element can figure out the format change
by itself.
2019-08-30 01:36:41 +09:00
Seungha Yang
f4f8941a91 nvdec: Add support 4:4:4 and 4:2:0 12bit decoding
Depending on GPU architecture, HEVC decoder can support
4:4:4 format up to 12 bitdepth. This commit covers VP9 4:2:0 12 bits
decoding also.
2019-08-29 13:39:59 +00:00
Seungha Yang
d0846f8eab nvdec: Port to GstCudaGraphicsResource
Make it possible to share registered graphics resource among nvidia encoders
and decoders.
2019-08-29 18:05:51 +09:00
Seungha Yang
8dc2b4a393 nvdec: Port to openGL PBO memory
For openGL interoperability, nvdec uses cuGraphicsGLRegisterImage API
which is to register openGL texture image.
Meanwhile nvenc uses cuGraphicsGLRegisterBuffer API to registure openGL buffer object.
That means two kinds of graphics resources are registered per memory
when nvdec/nvenc are configured at the same time.
The graphics resource registration brings possibly high overhead
so the registration should be performed only once per resource
from optimization point of view.
2019-08-29 18:04:33 +09:00
Seungha Yang
9bfd6d13e6 nvdec: Filter openGL API version to use
To ensure PBO buffer, openGL API >= 3 is required.
2019-08-29 18:04:29 +09:00
Seungha Yang
807e311ae8 nvdec: Always response QUERY_CONTEXT even if openGL is unavailable on the system
nvdec can response for the CUDA context type query regardless of openGL
availability.
2019-08-21 14:14:07 +09:00
Seungha Yang
4f60117db9 nvdec: Fix possible null object unref
gst_query_get_n_allocation_pools > 0 does not guarantee that
the N th internal array has GstBufferPool object. So users should
check the returned GstBufferPool object from
gst_query_parse_nth_allocation_pool.
2019-08-20 10:14:54 +09:00
Seungha Yang
eab564d857 nvcodec: Use default flag for CUDA stream creation
Since nvdec/nvenc engine is running on default stream,
non-default CUDA stream should be synchronized with default
stream eventually.
2019-08-19 07:13:26 +00:00
Seungha Yang
5615e9258f nvdec: Don't use default CUDA stream
Async CUDA operation with default stream (NULL CUstream) is not much
beneficial than blocking operation since all CUDA operations which belong
to the CUDA context will be synchronized with the default stream's operation.
Note that CUDA stream will share all resources of the corresponding CUDA context
but which can help parallel operation similar to the relation between thread and process
2019-08-19 01:18:52 +00:00
Seungha Yang
20d8f54e63 nvdec: Push/Pop CUDA context around library API call 2019-08-19 01:18:52 +00:00
Seungha Yang
f7b2b1b99d nvdec: Fix timestamp mismatch on draining frames
The internal decoding state must be GST_NVDEC_STATE_PARSE before
calling CuvidParseVideoData(). Otherwise, nvdec will be confused
on decode callback as if the frame is decoding only frame and
the input timestamp of corresponding frame will be ignored.
Eventually one decoded frame will have non-increased PTS.
2019-08-18 15:52:32 +09:00
Seungha Yang
b64733972e nvdec: Do not access nvdec object from destroy function of qdata
The destroy callback can be called just before the fìnalization of
GstMiniObject. So the nvdec object might be destroyed already.
Instead, store the GstCudaContext with increased ref to safely
unregister the CUDA resource.
2019-08-16 19:40:31 +09:00
Seungha Yang
9d0545d1a2 nvcodec: Wrap CUDA API return check with gst_cuda_result
The gst_cuda_result macro function is more helpful for debugging
than previous cuda_OK because gst_cuda_result prints the function
and line number. If the CUDA API return was not CUDA_SUCCESS,
gst_cuda_result will print WARNING level debug message with
error name, error text strings.
2019-08-07 00:59:36 +00:00
Seungha Yang
d69b590683 nvdec: Port to GstCUDAContext
... and drop CUvideoctxlock usage. The CUvideoctxlock basically
has the identical role of cuda context push/pop but nvdec specific
way. Since we can share the CUDA context among encoders and decoders,
use CUDA context directly for accessing GPU API.
2019-08-07 00:59:36 +00:00
Seungha Yang
f7f9f327cd nvdec: Respect upstream provided timestamp
Decoder sometimes reports nonincreasing timestamp.
Use input frame's timestamp like other decoder elements.
2019-08-05 20:32:39 +00:00
Seungha Yang
c99b160b50 nvdec: Use upstream framerate if possible
Encoded bitstream might not have valid framerate. If upstream
provided non-variable-framerate (i.e., fps_n > 0 and fps_d > 0)
use upstream framerate instead of parsed one.
2019-08-05 15:32:43 +00:00
Seungha Yang
f1cbab7cfd nvdec: Fix build warning error
gstnvdec.c:1222:3: error: implicit declaration of function ‘memset’ [-Werror=implicit-function-declaration]
   memset (&type_info, 0, sizeof (type_info));
   ^~~~~~
2019-07-31 15:36:04 +00:00
Seungha Yang
694f91da88 nvdec: Make OpenGL dependency optional
By adding system memory support for nvdec, both en/decoder
in the nvcodec plugin are able to be usable regardless of
OpenGL dependency. Besides, the direct use of system memory
might have less overhead than OpenGL memory depending on use cases.
(e.g., transcoding using S/W encoder)
2019-07-26 00:01:23 +00:00
Seungha Yang
733c109ce9 nvcodec: Clean up pointless return values around plugin init
Any plugin which returned FALSE from plugin_init will be blacklisted
so the plugin will be unusable even if an user install required runtime
dependency next time. So that's the reason why nvcodec returns TRUE always.

This commit is to remove possible misreading code.
2019-07-25 08:47:50 +00:00
Seungha Yang
e5a98cf9d8 nvdec: Add support for 10bits 4:2:0 decoding
This commit includes h265 main-10 profile support if the device can
decode it.

Note that since h264 10bits decoding is not supported by nvidia GPU for now,
the additional code path for h264 high-10 profile is a preparation for
the future Nvidia's enhancement.
2019-07-25 08:06:26 +00:00
Seungha Yang
d692350fc3 nvdec: Specify supported profiles of h264/h265 codec
See more details about supported formats at
nvidia codec sdk document "NVDEC_VideoDecoder_API_ProgGuide.pdf"
Table 1. Hardware Video Decoder Capabilities.

Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/issues/926
2019-07-25 08:06:26 +00:00
Seungha Yang
c8640e23f4 nvdec: Skip draining before creating internal parser
GstVideoDecoder::drain/flush can be called at very initial state
with stream-start and flush-stop event, respectively.
Draning with NULL CUvideoparser seems to unsafe and that eventually
failed to handle it.
2019-07-25 07:11:04 +00:00
Seungha Yang
a2ada54265 nvcodec: Keep requested rank for default device
Fix for default encoder and decoder element factory to make them have
higher rank than the others.
2019-07-23 10:28:52 +09:00
Seungha Yang
0239152bca nvdec: Create CUDA context with registered device id
Only the default device has been used by NVDEC so far.
This commit make it possible to use registered device id.
To simplify device id selection, GstNvDecCudaContext usage is removed.
2019-07-22 17:39:45 +00:00
Seungha Yang
1df2f13d0c nvdec: Register elements per device/codec with capability check
By this commit, each codec has its own element factory so the
nvdec element factory is removed. Also, if there are more than one device,
additional nvdec element factory will be created per
device like nvh264device{device-id}dec, so that the element factory
can expose the exact capability of the device for the codec.
2019-07-22 17:39:45 +00:00
Seungha Yang
48a6641717 nvdec: Fix video stuttering issue with VP9
Address nvidia driver specific behavior to avoid unexpected frame mismatch
between GStreamer and NVDEC.
2019-07-19 18:44:32 +09:00
Seungha Yang
8018fa2526 nvdec: Drop async queue and handle data on callback of CUvideoparser
Callbacks of CUvideoparser is called on the streaming thread.
So the use of async queue has no benefit.

Make control flow straightforward instead of long while/switch loop.
2019-07-19 18:44:32 +09:00
Seungha Yang
8753561015 nvdec: Port to color_{primaries,transfer,matrix}_to_iso
... and update the color information only when upstream was not provided
the information.
2019-07-17 06:34:21 +00:00
Seungha Yang
8862abd7c6 nvdec: Fix possible frame drop on EOS
On eos, baseclass videoencoder call finish() vfunc instead of drain()
2019-07-09 20:52:23 +09:00
Seungha Yang
c18fda03d9 nvdec,nvenc: Port to dynamic library loading
... and put them into new nvcodec plugin.

* nvcodec plugin
Now each nvenc and nvdec element is moved to be a part of nvcodec plugin
for better interoperability.
Additionally, cuda runtime API header dependencies
(i.e., cuda_runtime_api.h and cuda_gl_interop.h) are removed.
Note that cuda runtime APIs have prefix "cuda". Since 1.16 release with
Windows support, only "cuda.h" and "cudaGL.h" dependent symbols have
been used except for some defined types. However, those types could be
replaced with other types which were defined by "cuda.h".

* dynamic library loading
CUDA library will be opened with g_module_open() instead of build-time linking.
On Windows, nvcuda.dll is installed to system path by CUDA Toolkit
installer, and on *nix, user should ensure that libcuda.so.1 can be
loadable (i.e., via LD_LIBRARY_PATH or default dlopen path)
Therefore, NVIDIA_VIDEO_CODEC_SDK_PATH env build time dependency for Windows
is removed.
2019-07-08 10:37:46 +00:00
Renamed from sys/nvdec/gstnvdec.c (Browse further)